Numerical Tools for analyzing cell signalling circuits in normal and tumor cells

描述（由申请人提供）：用于分析正常和肿瘤细胞中细胞信号通路的数值工具该提案是分阶段STTR奖的第一部分，描述了软件工具的开发，工作流程和原理验证案例研究，用于细胞信号通路（以及类似的复杂生物过程）的详细动力学建模。建立真实的生物过程动力学模型的最关键步骤是优化模型预测与实验数据之间的匹配。实现这种优化需要有效的手段来探索替代途径拓扑结构（这会导致模型的结构）和“校准”模型的数据（这约束模型参数）。我们解决这两个问题，特别强调的算法和软件需要进行校准的高参数模型的基础上确定性常微分方程。这些模型的大小，实验数据的不完整性以及广泛的时间尺度使得有效的校准在理论和实践中都极具挑战性，使得该提案在生物学和应用数学方面都具有创新性和重大意义。（数值）在化学建模，燃烧工程和数值模拟和两个系统生物学小组（Sorger和Gunawardena）分别在实验生物学和计算机科学的专业知识深厚的专业知识。第一阶段的三个具体目标包括：（1）开发基于Lisp的小B编程环境，作为一种手段，以快速修订的语言表示生物途径和多组分蛋白质复合物，（2）优化用于校准生物模型的数值分析工具，采用复杂的工作流程和基于伴随的灵敏度分析和全局参数估计的新算法（3）系统药理学和药物发现中验证模型和将关键发现应用于重大问题的方法。

DESCRIPTION (provided by applicant): Numerical Tools for analyzing cell signaling circuits in normal and tumor cells The proposal, part I of a phased STTR award, describes the development of software tools, workflow and proof-of-principle case studies for detailed kinetic modeling of cell signaling pathways (and similarly complex biological processes). The most critical step in building kinetic models of real biological processes is optimizing the match between model prediction and experimental data. Achieving this optimization requires efficient means to explore alternative pathway topologies (which give rise to the structure of the model) and to "calibrate" models to data (which constrains model parameters). We address both of these problems with particular emphasis on the algorithms and software needed to perform calibration of highly parametric models based on deterministic ordinary differential equations. The size of these models, the incompleteness of experimental data, and the wide range of time scales makes effective extremely calibration challenging both in theory and in practice, making the proposal innovative and significant with respect to both its biology and its applied mathematics The project is a close collaboration between a software company (Numerical) with deep expertise in chemical modeling, combustion engineering and numerical simulation and two systems biology groups (Sorger and Gunawardena) with expertise in experimental biology and computer science respectively. Three phase I specific aims include (1) Development of the Lisp-based little b programming environment as a means to represents biological pathways and multi-component protein complexes in a rapidly- revised, reusable form suitable (2) Optimizing numerical analysis tools for calibration of biological models with a sophisticated workflow and new algorithms for adjoint-based sensitivity analysis and global parameter estimation (3) Methods for validating models and applying key findings to significant problems in systems pharmacology and drug discovery.